|
On November 27 2014 01:18 Cyro wrote:Show nested quote +On November 27 2014 01:16 Ropid wrote:
I only know about going through all devices and their drivers and seeing if people report issues online for one of them.
Did something get updated recently, like did it only start with the new NVIDIA drivers for example? Windows Update could also have secretly installed a new driver for a device.
It could also be an update to Windows internals like kernel and the graphics stuff as I think Microsoft regularly ports things from Windows 8 back to 7. You could try to see if people post in forums about something like that by having Google only report search result from the past week or past month.
About strange reasons, I wonder if something like the C: drive having issues could cause this. well it dropped from 5k to 25 (dropped by 99.5%..) as soon as i disabled those c-states etc. I'm not sure what could have changed because i didn't check it highly regularly, but i'm using a new mouse and a new display driver - that's the ONLY thing plugged into USB on my whole system (ps/2 keyboard, nothing else) and those two were very high on the list on the second pic i posted on first spoiler while i'm posting here: My vg248qe got another messed up pixel, stuck pixel is i think the right term. It appears pink when you put white behind it, changes a bit with different colors and is not really visible (or not visible at all) when it's supposed to display black. There's two of them now (both were not there when i first got monitor) within about a 1/5'th slice of the monitor near the right side, is that anything to worry about? One was a very minor annoyance to me, but i don't want another 5 to show up at this rate considering i didn't have the monitor for very long and i kinda wanted to keep it for something like 5+ years (if not for my usage, maybe someone elses close to me like my dad or cousin for example)
There's Youtube videos that are supposedly helping with stuck pixels. It's flashing colors for hours and you are supposed to run that full screen. I have no idea if that works often or rarely or never. 
About your power saving question, most of the power saving is done by that C6/C7 stuff you have now disabled. That feature can shut down power to individual cores and can shut down more parts of the CPU if all the cores are in the same C6/C7 state. The power saving is supposed to come from things simply being off.
The C1E and EIST stuff is about reducing speed. On stock settings, this will then cause the voltage to drop (no idea what's going on with Haswell there if you overclock). The power saving is supposed to come from the reduced voltage.
+ Show Spoiler +That C1E stuff is about reducing speed (so that voltage can drop), when a core is at low load. There's an instruction named "halt" that will make a core stop and only continue if there's an interrupt signal from the outside coming in. After the "halt", the core is in the C1 state. The C1E feature tries to improve power saving by doing things to the clock and power. When an interrupt happens, the core can then only start working at 1600 MHz (or 800 MHz on Haswell?) until the voltage ramps up again. If it's done processing the interrupt and hits another "halt" instruction fast enough, it will never get to ramp up speed and voltage.
This will cause DPC latency on the desktop to look a bit bad, simply because stuff gets processed slower. This should only be the case on the desktop because there's only one speed for all cores. When a game runs, there will always be one core running at full speed and that means all cores run at full speed.
That C1E speed reducing stuff is automatic. EIST is the feature that allows Windows to put a limit for max speed. Windows looks at the CPU load for the cores to decide on what to do about that limit. If the way overclocking is working on your board and CPU means that voltage will be low when running at lower speeds, this is the important setting for you if you worry about high voltage being used too much on the desktop. C1E alone won't keep things from often hitting full speed when only clicking around on the desktop.
All of that speed reducing stuff is a bit disappointing with regards to power saving in comparison to C6/C7. If a core executed that "halt" instruction and is in C1, it will by itself go into C6 after a while. If you disable C1E and keep C6 enabled, this means when an interrupt happens the core will wake up with full voltage and full speed. It will use more power, but it will also process the interrupt faster than it would have at 1600 MHz and can then go back to sleep faster.
When you look at the task manager and the CPU load graphs, that's basically showing how much the CPU cores are stopped. The cores can switch between on and off in microseconds and if it shows 5% load on the desktop, this means cores are not running 95% of the time.
|
United Kingdom20345 Posts
|
On November 27 2014 02:28 Ropid wrote:Show nested quote +On November 27 2014 01:18 Cyro wrote:On November 27 2014 01:16 Ropid wrote:
I only know about going through all devices and their drivers and seeing if people report issues online for one of them.
Did something get updated recently, like did it only start with the new NVIDIA drivers for example? Windows Update could also have secretly installed a new driver for a device.
It could also be an update to Windows internals like kernel and the graphics stuff as I think Microsoft regularly ports things from Windows 8 back to 7. You could try to see if people post in forums about something like that by having Google only report search result from the past week or past month.
About strange reasons, I wonder if something like the C: drive having issues could cause this. well it dropped from 5k to 25 (dropped by 99.5%..) as soon as i disabled those c-states etc. I'm not sure what could have changed because i didn't check it highly regularly, but i'm using a new mouse and a new display driver - that's the ONLY thing plugged into USB on my whole system (ps/2 keyboard, nothing else) and those two were very high on the list on the second pic i posted on first spoiler while i'm posting here: My vg248qe got another messed up pixel, stuck pixel is i think the right term. It appears pink when you put white behind it, changes a bit with different colors and is not really visible (or not visible at all) when it's supposed to display black. There's two of them now (both were not there when i first got monitor) within about a 1/5'th slice of the monitor near the right side, is that anything to worry about? One was a very minor annoyance to me, but i don't want another 5 to show up at this rate considering i didn't have the monitor for very long and i kinda wanted to keep it for something like 5+ years (if not for my usage, maybe someone elses close to me like my dad or cousin for example) There's Youtube videos that are supposedly helping with stuck pixels. It's flashing colors for hours and you are supposed to run that full screen. I have no idea if that works often or rarely or never. About your power saving question, most of the power saving is done by that C6/C7 stuff you have now disabled. That feature can shut down power to individual cores and can shut down more parts of the CPU if all the cores are in the same C6/C7 state. The power saving is supposed to come from things simply being off. The C1E and EIST stuff is about reducing speed. On stock settings, this will then cause the voltage to drop (no idea what's going on with Haswell there if you overclock). The power saving is supposed to come from the reduced voltage. + Show Spoiler +That C1E stuff is about reducing speed (so that voltage can drop), when a core is at low load. There's an instruction named "halt" that will make a core stop and only continue if there's an interrupt signal from the outside coming in. After the "halt", the core is in the C1 state. The C1E feature tries to improve power saving by doing things to the clock and power. When an interrupt happens, the core can then only start working at 1600 MHz (or 800 MHz on Haswell?) until the voltage ramps up again. If it's done processing the interrupt and hits another "halt" instruction fast enough, it will never get to ramp up speed and voltage.
This will cause DPC latency on the desktop to look a bit bad, simply because stuff gets processed slower. This should only be the case on the desktop because there's only one speed for all cores. When a game runs, there will always be one core running at full speed and that means all cores run at full speed.
That C1E speed reducing stuff is automatic. EIST is the feature that allows Windows to put a limit for max speed. Windows looks at the CPU load for the cores to decide on what to do about that limit. If the way overclocking is working on your board and CPU means that voltage will be low when running at lower speeds, this is the important setting for you if you worry about high voltage being used too much on the desktop. C1E alone won't keep things from often hitting full speed when only clicking around on the desktop.
All of that speed reducing stuff is a bit disappointing with regards to power saving in comparison to C6/C7. If a core executed that "halt" instruction and is in C1, it will by itself go into C6 after a while. If you disable C1E and keep C6 enabled, this means when an interrupt happens the core will wake up with full voltage and full speed. It will use more power, but it will also process the interrupt faster than it would have at 1600 MHz and can then go back to sleep faster.
When you look at the task manager and the CPU load graphs, that's basically showing how much the CPU cores are stopped. The cores can switch between on and off in microseconds and if it shows 5% load on the desktop, this means cores are not running 95% of the time.
I'm not sure if that program actually helps jog stuck pixels or just makes it easier to identify them - I've used it once to identify dead pixels. Still, worth a shot.
|
For over a year I've been playing on NA from EU with no notable delay. I have experienced absolutely no lag or delay on NA whatsoever.
And now... within the last month or something like that, I am having delay every game. It's not much compared to Kr server, but it is enough for me not wanting to play. I can't blink stalker micro properly.
Have blizzard changed something with the servers recently. Or what else can cause this. My internet is exactly the same.
|
edit: FAO Mozdk Denmark. December 01 2014 20:57
If you run a tracert command, you'll be able to identify where in the series of hops from your computer, through your ISP to blizz's servers the problem may lie. Spotting where the problem lie is usually quite obvious as the ping for one hop may be significantly higher that the others, or perhaps times out altogether.
I'm assuming you're running a version of windows here.
Had a quick gander and found the following as the list of SC2 servers,
The Americas 12.129.202.154 ----- 63.240.104.93
Europe 195.12.232.102 ----- 195.12.232.70
Korea and Taiwan 211.234.110.1
Southeast Asia 202.9.67.254
So if you wanted to check the NA ones you'd have to run 2 commands, one for each IP.
Click START, type in cmd right click the cmd.exe and run as administrator, which will pop open the command window, and type in;
tracert 12.129.202.154 > C:\Users\Rob\Desktop\12-129.txt
tracert 63.240.104.93 > C:\Users\Rob\Desktop\63-240.txt
You will have to change Rob with your username on your computer, the > is a pipe, so it's outputting the command to a text file it will place on your desktop, with a simple name tag to point which is for which IP. You can have it place the file anywhere you want, just alter the C:\Users\Rob\Desktop\ part.
Be patient and wait till it ends.
If for example you have a huge spike in ping say on the second hop down, it's possible this is still in your ISP's system, so you may be able to contact them to get advice or assistance. If any spike in ping happens near the end, then this is down the other end of the line and near to Blizzards servers, perhaps within their own infrastructure, again, you may be able to get help for this from Blizzard, if it's in the middle section of hops that's a bit of a no mans land in most cases, and may prove hard to resolve.
|
ps. this is all assuming that it's an ISP related issue - something to do with the routing from A to B. You have good antivirus I assume? I've read that malware or virii can cause drastic connection issues. What about WiFi or ethernet connection from your computer to your router, ethernet is better, period. If you're forced to use Wifi, you can download apps for your smart phone which can look at the channels being used in the area and suggest potentially better ones, though I am not overly sure how effective this is, I am led to belive that most modern WiFi routers change channels as they see best. WiFi analyzer is one such free app on android however.
|
Can an internal SATA non system hard drive cause an auto shutdown/blue screen in windows 7?
My computer crashes, but it seems to be only when I'm pulling data from that extra drive.
|
On December 02 2014 10:38 LeperKahn wrote:
Can an internal SATA non system hard drive cause an auto shutdown/blue screen in windows 7?
My computer crashes, but it seems to be only when I'm pulling data from that extra drive.
Run a diagnostic utility on the drive, would be step one I guess. Many manufactureres produce their own. CrystalDiskInfo is a popular freeware one. Search something along the lines of 'hard disk drive diagnostic utility freeware' to find others.
|
|
|
I am having a sudden problem with my GTX 770 (PNY VCGGTX7702XPB) after restarting my computer for the first time in a while. In both BF4 and SC2, I was getting a crash and error message: "DirectX function "GetDeviceRemovedReason" failed with DXGI_ERROR_DEVICE_HUNG". I updated the graphics drivers and got the same problem (although before I could tell the driver update was completed, the monitor went black and I had to restart). The last time this happened in game it crashed my card completely and now, I can't get any signal to the monitor on start up. It doesn't even display the motherboard post screen. Dafuq?
EDIT: connected my monitor to the onboard video port, changed display device in BIOS to onboard, and can see motherboard post and get to the desktop. The GTX 770 is being recognized in the device manager it just says "Standard VGA Adapter" and it also not being detected by speedfan. But is this just because I changed the display device to onboard in the BIOS?
|
|
|
United Kingdom20345 Posts
On December 02 2014 13:48 Doodsmack wrote:
I am having a sudden problem with my GTX 770 (PNY VCGGTX7702XPB) after restarting my computer for the first time in a while. In both BF4 and SC2, I was getting a crash and error message: "DirectX function "GetDeviceRemovedReason" failed with DXGI_ERROR_DEVICE_HUNG". I updated the graphics drivers and got the same problem (although before I could tell the driver update was completed, the monitor went black and I had to restart). The last time this happened in game it crashed my card completely and now, I can't get any signal to the monitor on start up. It doesn't even display the motherboard post screen. Dafuq?
EDIT: connected my monitor to the onboard video port, changed display device in BIOS to onboard, and can see motherboard post and get to the desktop. The GTX 770 is being recognized in the device manager it just says "Standard VGA Adapter" and it also not being detected by speedfan. But is this just because I changed the display device to onboard in the BIOS?
It might be dead/dieing
It explains the error in the message, that one happens all the time because blizzard patches different regions at different times. If you updated your game to play on KR for example and the EU/US servers did not get that update yet, you would only be able to play on KR until they did
|
NA is working now. Still can't go on EU
|
On December 03 2014 06:57 Mozdk wrote:
NA is working now. Still can't go on EU
Any1 want to give responds to Whether or not they are experiencing this too?!?
|
United Kingdom20345 Posts
It always happens when you patch the game to a version that not all servers are updated to. IIRC korea patches monday, US tuesday, europe wednesday (in the mid-morning)
|
On December 03 2014 09:30 Cyro wrote:
It always happens when you patch the game to a version that not all servers are updated to. IIRC korea patches monday, US tuesday, europe wednesday (in the mid-morning)
Then why have I been able to play for 4 years, every day, without ever seeing this? I've always been able to play right up until the scheduled maintenance. Which is in 1.5 hours.... And I haven't been able to play for 4 ish hours...
|
United Kingdom20345 Posts
On December 03 2014 09:37 Mozdk wrote:Show nested quote +On December 03 2014 09:30 Cyro wrote:
It always happens when you patch the game to a version that not all servers are updated to. IIRC korea patches monday, US tuesday, europe wednesday (in the mid-morning) Then why have I been able to play for 4 years, every day, without ever seeing this? I've always been able to play right up until the scheduled maintenance. Which is in 1.5 hours.... And I haven't been able to play for 4 ish hours...
It would work if you didn't update your game to a version that hadn't been pushed to US/EU servers yet. It's only a problem if you play Korean/US server on patch day, then it's resolved within 48/24 hours or whenever you reinstall/downgrade game client
|
So my computer is finally dying after a really good run of like 6+ years on some of the parts and I'm wondering if anyone has shipped internationally before and if it's worth? I'm in California right now but in ~1 month I will be back home in GA. I have a friend's laptop that would tide me over till then in case it actually dies. It's the PSU I think. I'm required to unplug my computer for the workday [ military ] and it takes 5 mins to boot up after being plugged in and the power button on my case being unresponsive.
I would save my hard drives and dump the rest, since its really old parts. I'd probably keep my monitor as well since it's pretty recent [ Dell U2312HM ]. If I decide to build a desktop it'd be from the US <> Korea. There's really old threads out there and mostly using UPS or USPS but just asking in case there's better alternatives.
I'm pretty sure the brand availability in Korea is less plus it's not that much cheaper to build one there than to build one here and ship it. And although it shouldn't make a difference, I'm more comfortable putting it together myself from parts ordered within the US. I'm pretty averse to laptops in general so unless there's some ridiculous bang for buck laptop out there ~$1000 I won't consider laptops. So yeah, build and ship? and with what company? It's really weird because I had my family mail my PC from GA<>CA as insured as possible and it cost like $300 at the time which I'm sure they ripped off my family member who sent it. It's a full size tower. But from reading online it seems to be way less than $300 to ship it safe enough without all the bells and whistles. Would appreciate a ballpark estimate from here to Korea or even domestically if its worth it to replace the PSU and keep the system going. I am interested in a brand new system though.
|
My mechanical hard drive just started making a noise. I would describe the noise as the sound of something spinning quickly every four seconds, for about a second, and then stopping, and repeating. The noise is kind of low pitch because of the way it is vibrating. Would something be causing this or is my seagate harddrive failing? 
|
Durak, seagate make a diagnostic tool.
http://www.seagate.com/gb/en/support/downloads/seatools/
Download that, run it, see what it says. Either way if your sure it is the hard drive I'm thinking it cant be a good sign. If this were me, I'd not use it until I had a replacement, pull as much info I could off it, wipe the drive, and bin it.
|
|
|
|
|
|
EPT
